Circulating Levels of Antioxidant Vitamins Correlate with Better Lung Function and Reduced Exposure to Ambient Pollution

To the Editor:

Particulate matter (PM) is possibly the ambient air pollution (AAP) that has the greatest effect on human health. Several studies consistently report an inverse association between PM exposure and lung function (e.g., FEV₁ and FVC) (1–6) and accelerated progression of chronic obstructive pulmonary disease (7) in adults. Furthermore, improved air quality associates with attenuated age-related decline in lung function (8).

Although several plausible mechanistic pathways have been described, the underlying mechanisms linking AAP and lung function have not been fully characterized.

High-throughput metabolomics approaches allow for an extensive set of small molecules to be measured in biological fluids. These metabolites represent pathways that reflect physiological functions, allowing for the potential identification of biomarkers (9).

To test the molecular links between lung function and AAP, we investigated the association between lung function and metabolomic parameters and between the same metabolites and exposure to PM (with aerodynamic diameter ≤ 10 μm [PM₁₀] and ≤ 2.5 μm [PM_2.5]) in the TwinsUK cohort.

Nontargeted metabolomic profiling using the Metabolon platform (280 metabolites) was performed in 5,519 fasting individuals from the TwinsUK cohort who also had undergone spirometry (Vitalograph model 2150, Buckingham, England), as described previously (10). A subset of 523 TwinsUK participants also had estimates of long-term exposure to PM at participants’ postal code residences, derived from a 20 × 20 m dispersion model for London. All study participants completed a medical history and lifestyle questionnaire, including questions on vitamin supplementation (see online supplement for details) (11).

The study was approved by St. Thomas' Hospital Research Ethics Committee, and all participants provided informed written consent. TwinsUK data are publicly available on request on the department website (http://www.twinsuk.ac.uk/data-access/accessmanagement/).

We inverse-normalized the metabolomics data and excluded metabolic traits with more than 20% missing values. Metabolite associations with FEV₁ and FVC were assessed by random intercept linear regressions adjusted for age, sex, body mass index, height, metabolite batch, and family relatedness. We adjusted for multiple testing, using Bonferroni correction, resulting in a significant threshold of 8.0 × 10⁻⁵ (=0.05/[280 metabolites × 2 traits]). The metabolites associated with FEV₁ and/or FVC were tested for correlation with PM_2.5 and PM₁₀ after adjustment for covariates and multiple testing.

The descriptive characteristics of the study population are shown in Table E1 in the online supplement. After adjustment for age, sex, body mass index, height, and family relatedness, exposure to PM_2.5 was seen to correlate negatively with both FEV₁ (β = −0.03; 95% confidence interval [CI], −0.04 to −0.01; P = 0.001; Figure 1) and FVC (β = −0.02; 95% CI, −0.04 to −0.01; P = 0.01). Similar results were found for PM₁₀ (FEV₁: β = −0.02 [95% CI, −0.04 to −0.01; P = 0.002]; FVC: β = −0.02 [95% CI, −0.03 to −0.001; P = 0.02]).

Figure 1.

Associations among lung function, exposure to particulate matter with aerodynamic diameter ≤ 2.5 μm (PM_2.5), and α-tocopherol circulating levels. The minimum and maximum values for each quartile of PM_2.5 exposure (A and B) and FEV₁ (C) are shown. (A) Mean normalized FEV₁ adjusted for age, height, and sex among quartiles of PM_2.5 exposure. (B) Mean normalized α-tocopherol levels adjusted for age, sex, and batch among quartiles of PM_2.5 exposure. (C) Mean normalized α-tocopherol levels adjusted for age, sex, and batch among quartiles of FEV₁. *FEV₁ values were adjusted for age, height, sex, and body mass index and are normalized and standardized to have a mean = 0, SD = 1.

FEV₁ correlates with 18 metabolites (Table E2), which fall into three principal classes: eight amino acids primarily involved in glycine, serine, and threonine metabolism; four cofactors and vitamins; and three lipids. There is also one carbohydrate, one nucleotide, and one xenobiotic.

The metabolomics analysis for FVC revealed 13 significantly associated metabolites, 10 of which were also identified for FEV₁ (Table E2). The associated metabolites were then tested for correlation with PM_2.5 and PM₁₀ in a subset of 532 TwinsUK individuals living in the Greater London area.

Of the 21 metabolites associated with lung function, eight were also significantly associated with both PM_2.5 and PM₁₀ (Table 1). Among the eight metabolites identified, four are amino acids, one is a carbohydrate (glycerate), one is a salt (benzoate), and two are cofactors and vitamins; namely, α-tocopherol and threonate. In all eight instances, a higher exposure to PM correlates with lower levels of the metabolite and a lower FEV₁ value (Table 1).

Table 1.

Metabolomic Associations with Airborne Particulate Matter and Lung Function and Their Association with CRP

			PM2.5		PM10		FEV1		FVC		CRP*
Metabolite	Super-p	Sub-p	β (SE)	P Value	β (SE)	P Value	β (SE)	P Value	β (SE)	P Value	β (SE)	P Value
Asparagine	a-a	Alanine and aspartate metabolism	−0.1 (0.02)	5.68 × 10−8	−0.05 (0.02)	9.09 × 10−4	0.03 (0.01)	2.46 × 10−5	0.02 (0.01)	6.09 × 10−3	−0.05 (0.01)	1.10 × 10−5
Glycine	a-a	Glycine, serine, and threonine metabolism	−0.11 (0.02)	9.06 × 10−9	−0.07 (0.02)	3.52 × 10−5	0.03 (0.01)	9.28 × 10−6	0.03 (0.01)	2.03 × 10−6	−0.08 (0.01)	1.20 × 10−12
N-acetylglycine	a-a	Glycine, serine, and threonine metabolism	−0.12 (0.02)	9.11 × 10−10	−0.08 (0.02)	4.63 × 10−7	0.03 (0.01)	6.87 × 10−6	0.04 (0.01)	3.93 × 10−9	−0.04 (0.01)	8.20 × 10−5
Serine	a-a	Glycine, serine, and threonine metabolism	−0.12 (0.02)	3.97 × 10−9	−0.07 (0.02)	4.61 × 10−5	0.03 (0.01)	3.07 × 10−5	0.03 (0.01)	3.12 × 10−4	−0.07 (0.01)	2.20 × 1
Glycerate	ch	Glycolysis, gluconeogenesis, pyruvate metabolism	−0.14 (0.02)	9.47 × 10−11	−0.08 (0.02)	8.28 × 10−7	0.03 (0.01)	2.06 × 10−7	0.03 (0.01)	3.00 × 10−4	−0.05 (0.01)	3.20 × 10−5
Threonate	c&v	Ascorbate and aldarate metabolism	−0.12 (0.02)	8.96 × 10−9	−0.08 (0.02)	1.99 × 10−5	0.03 (0.01)	4.80 × 10−6	0.02 (0.01)	1.29 × 10−3	−0.04 (0.01)	5.30 × 10−5
α-Tocopherol	c&v	Tocopherol metabolism	−0.17 (0.02)	4.92 × 10−13	−0.12 (0.02)	3.48 × 10−10	0.04 (0.01)	5.31 × 10−7	0.04 (0.01)	3.57 × 10−6	−0.03 (0.01)	6.30 × 10−3
Benzoate	x	Benzoate metabolism	−0.11 (0.02)	6.02 × 10−6	−0.07 (0.02)	3.13 × 10−4	0.03 (0.01)	1.84 × 10−5	0.03 (0.01)	1.67 × 10−5	−0.01 (0.01)	1.80 × 10−1

Definition of abbreviations: a-a = amino-acid; ch = carbohydrate; CRP = C-reactive protein; c&v = cofactor and vitamin; PM_2.5 = particulate matter with aerodynamic diameter ≤ 2.5 μm; PM₁₀ = particulate matter with aerodynamic diameter ≤ 10 μm; x = xenobiotics.

^*Log of CRP circulating levels. CRP was also tested for association with lung function and particulate matter. None of these comparisons reached nominal statistical significance (P < 0.05) after adjusting for the same covariates as for the metabolomics panel. The regression coefficients observed with logCRP were PM₁₀ β (SE) = 0.009 (0.007), P = 0.23; PM_2.5 β (SE) = 0.013 (0.009), P = 0.18; FEV₁ β (SE) = −0.151 (0.253), P = 0.565; FVC β (SE) = −0.073 (0.360), P = 0.843.

Seven of the eight metabolites identified correlate negatively with circulating levels of C-reactive protein, a marker of generalized inflammation (Table 1). The amino acids identified are all highly correlated with each other (Table E3), and in particular, glycine has been linked in the literature to pulmonary inflammation (see online supplement).

The strongest association both with PM_2.5 and FEV₁ was seen with vitamin E (Figures 1B and C). We also identified threonate, which is the major metabolite of ascorbic acid (vitamin C, a water soluble antioxidant) produced under oxidative conditions (12). We thus investigated use of vitamin supplements in study participants. We found positive significant correlations between use of vitamin supplements with both α-tocopherol (β = 0.096; SE = 0.05; P = 0.045) and threonate normalized levels (β = 0.19; SE = 0.06; P = 0.001). However, we found no correlation between use of vitamin supplements and PM_2.5 (β = 0.02; SE = 0.11; P = 0.86) and PM₁₀ (β = =0.03; SE = 0.01; P = 0.79).

In conclusion, circulating levels of eight metabolites are significantly correlated with both exposure to AAP and lung function. The strongest association both with PM_2.5 and FEV₁ was with α-tocopherol levels: individuals with a higher PM_2.5 exposure have significantly lower levels of α-tocopherol and also have lower lung function. To our knowledge, this is the first report of significant association between α-tocopherol levels and PM_2.5 exposure in the general population. This is consistent with previous literature reports indicating that antioxidants, particularly α-tocopherol (but not γ-tocopherol), result in improved lung function (13), as well as with the extensive body of evidence indicating that lower α-tocopherol levels are observed in lung challenges such as asthma (see Discussion in the online supplement).

α-Tocopherol is a biologically active form of the fat-soluble antioxidant vitamin E and also regulates gene expression (14). Supplementation with vitamin E reduces the damage to lung function caused by AAP in children (15). Circulating levels of α-tocopherol have also been shown to correlate positively with lung function in adults (13).

We note some study limitations: We could not access an independent population with PM and metabolomic data on which to confirm these results. Given the cross-sectional nature of the data, we were unable to make causal inference. However, in these data, use of vitamin supplements is positively correlated with circulating levels of both α-tocopherol and threonate, and as expected, there is no relationship between PM exposure and use of vitamins. Taken together, these findings suggest that subjects with lower levels of α-tocopherol are at greater risk of losing FEV₁ when exposed to urban PM. If this is indeed the case, the data may indicate that individuals with a high exposure to AAP would benefit from an optimal antioxidant status.